Owli-AI Custom GPT
SeeForMe
Structured image description, orientation, and OCR in ChatGPT for blind and severely visually impaired people.
Notice: this page was machine-translated and is pending editorial review.
- ChatGPT
- Low Vision
- OCR
Quick start
Structured image description, orientation, and OCR in ChatGPT for blind and severely visually impaired people.
- Open: Open the GPT via the ChatGPT link.
- Add context: Briefly describe your situation so the GPT can respond in a focused way.
- Ask a question: Start with a specific question or task.
Example prompts
- Describe the scene in 5 points. Use clock-face directions (12 o'clock, 3 o'clock) and rough distances.
- Read all text in the image (OCR) and output it cleanly formatted (keep lines/paragraphs).
- Are there possible obstacles or hazards? Name them briefly first, then provide details.
Typical questions
- Which objects are within reach, and where are they relative to each other (for example 'in front of you', 'lower left', '12 o'clock')?
- Give me short orientation guidance: what is the safest route/step from my current position?
- Summarize the image once as a short cue (1 sentence) and once as a detailed cue (3-5 sentences).
Limits and privacy
Requires a ChatGPT account.